Hacker News with Generative AI: Computer Science

Uninitialized garbage on ia64 can be deadly (2004) (microsoft.com)
On Friday, we talked about some of the bad things that can happen if you call a function with the wrong signature.
Look Ma, No Bubbles: Designing a Low-Latency Megakernel for Llama-1B (hazyresearch.stanford.edu)
There are some applications that benefit from running LLMs really, really fast. This low-latency regime encompasses applications like chatbots and human-in-the-loop workflows, where users care a lot about seeing responses come back immediately.
10biForthOS: A full 8086 OS in 46 bytes (git.sr.ht)
Grid-Free Approach to Partial Differential Equations on Volumetric Domains [pdf] (rohansawhney.io)
Grammars of Formal Uncertainty (arxiv.org)
Large language models (LLMs) show remarkable promise for democratizing automated reasoning by generating formal specifications.
Year 2038 Problem (wikipedia.org)
Computer Science major has seventh highest unemployment rate (newsweek.com)
Despite routinely ranking as one of the most popular majors for college students and recent graduates, computer science has one of the highest unemployment rates across all areas of study.
As a programmer, get your math sorted (csprimer.in)
Master the mathematical foundations essential for computer science.
You could have invented Transformers (gwern.net)
‘You Could Have Invented Transformers’ tutorial proposal
Open Source Society University – Path to a free self-taught education in CS (github.com/ossu)
The OSSU curriculum is a complete education in computer science using online materials.
20 years of NetBSD code bloat (blogspot.com)
I started with NetBSD in the mid-nineties, on a Sun SPARC ELC with 32Mbyte of memory, where I used GCC and Emacs on X11 with FVWM as the window manager. I'm still using GCC, Emacs, and FVWM with the same configuration files (updated for pointless changes in Emacs and FVWM), but I now need much more memory and CPU performance... I thought it would be interesting to investigate why.
Domain Theory Lecture Notes (liamoc.net)
The Verse Calculus: A Core Calculus for Functional Logic Programming [pdf] (peytonjones.org)
Alan Turing papers saved from shredder could fetch £150k (theguardian.com)
Widely considered the father of theoretical computer science, Alan Turing’s influence on modern life continues to be felt in the age of artificial intelligence. But despite this legacy, a cache of his most important papers was nearly shredded – only to be saved at the last minute when their significance was recognised at a family event.
Model-Based Machine Learning (2023) (mbmlbook.com)
Recreating a video from raw captured network traffic (2023) (thesloth.me)
A few days ago i stumbled upon the idea to recreate a video from traffic captured in PCAP file. It may be a common knowledge that this is possible, but it felt like the coolest thing i’ve done in a while. If you’re interested on how it was done continue reading.
CRDTs #2: Turtles All the Way Down (jhellerstein.github.io)
Modern distributed systems often seem to rest on an stack of turtles. For every guarantee we make, we seem to rely on a lower-layer assumption. Eventually we're left wondering: what is at the bottom?
New method for creating large 3D models of urban areas is faster and cheaper (techxplore.com)
A research team led by Waterloo Engineering has developed a faster, cheaper way to create large-scale, three-dimensional (3D) computer models of urban areas, technology that could impact fields including urban planning, architectural design and filmmaking.
The Next Abstraction (substack.com)
Study shows vision-language models can't handle queries with negation words (news.mit.edu)
Words like “no” and “not” can cause this popular class of AI models to fail unexpectedly in high-stakes settings, such as medical diagnosis.
Ask HN: Places in the UK / Europe Related to computers (ycombinator.com)
I’m interested in visiting some historic or special places related to this field as a way of rejuvenating my passion in the field again.
Dijkstra on Ada (wordpress.com)
Dijkstra was one of the reviewers of the fours language proposals which eventually lead to the language Ada: red, green, blue, and yellow (the winning proposal was green). Here are some quotes from each of the reviews:
CRDTs: Pros and Cons (Lattices and Lettuces?) (jhellerstein.github.io)
Over the next few days, I'm going to post a number of observations about CRDTs: Convergent Replicated Data Types. These are data structures that aspire to help us with coordination-free distributed programming, a topic that interests me a lot. How can developers (or languages/compilers) deliver distributed programs that are safe or correct in important ways, without employing expensive mechanisms for coordination that make the global cloud run as slowly as a sequential computer?
For algorithms, a little memory outweighs a lot of time (quantamagazine.org)
One computer scientist’s “stunning” proof is the first progress in 50 years on one of the most famous questions in computer science.
Everyone gets bidirectional BFS wrong (2024) (zdimension.fr)
People really need to stop blindly copying code from the Internet.
Alan Turing papers saved from shredder could fetch £150k (theguardian.com)
Widely considered the father of theoretical computer science, Alan Turing’s influence on modern life continues to be felt in the age of artificial intelligence. But despite this legacy, a cache of his most important papers was nearly shredded – only to be saved at the last minute when their significance was recognised at a family event.
Memory Consistency Models: A Tutorial (jamesbornholt.com)
There are, of course, only two hard things in computer science: cache invalidation, naming things, and off-by-one errors. But there is another hard problem lurking amongst the tall weeds of computer science: seeing things in order. Whether it be sorting, un-sorting, or tweeting, seeing things in order is a challenge for the ages.
Definitive solution to the hardest problem in computing: P = NP? (researchgate.net)
Why Concatenative Programming Matters (2012) (blogspot.com)
Concatenative programming is so called because it uses function composition instead of function application—a non-concatenative language is thus called applicative.
Bits with Soul (cam.ac.uk)
Watch Simon’s lecture